embedding-visualizer 0.1.0__tar.gz

embedding_visualizer-0.1.0/.github/workflows/workflow.yaml

@@ -0,0 +1,20 @@
+name: Publish to PyPI
+
+on:
+  push:
+    branches:
+      - main
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    permissions:
+      id-token: write
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+      - run: pip install build
+      - run: python -m build
+      - uses: pypa/gh-action-pypi-publish@release/v1

embedding_visualizer-0.1.0/.gitignore

@@ -0,0 +1,8 @@
+__pycache__/
+*.pyc
+*.egg-info/
+dist/
+build/
+.venv/
+
+examples/*.html

embedding_visualizer-0.1.0/PKG-INFO

@@ -0,0 +1,11 @@
+Metadata-Version: 2.4
+Name: embedding-visualizer
+Version: 0.1.0
+Summary: Visualize text embeddings with interactive plots
+Requires-Python: >=3.10
+Requires-Dist: nbformat>=4.2.0
+Requires-Dist: numpy>=1.24.0
+Requires-Dist: openai>=1.0.0
+Requires-Dist: plotly>=5.18.0
+Requires-Dist: scikit-learn>=1.3.0
+Requires-Dist: tiktoken>=0.5.0

embedding_visualizer-0.1.0/README.md

@@ -0,0 +1,72 @@
+# embedding-visualizer
+
+Interactive 2D visualization of text embeddings using OpenAI's embedding API and Plotly.
+
+## Install
+
+```bash
+cd embedding_visualizer
+uv sync
+```
+
+Requires an `OPENAI_API_KEY` environment variable.
+
+## Usage
+
+```python
+from embedding_visualizer import visualize_embeddings, PrincipalComponent, TextEmbedding
+
+docs = [
+    {
+        "text": "This text will be embedded",
+        "label": "optional, groups points in the legend",
+        "color": "optional, color for the point/label group",
+        "line-id": "optional, connects points with the same id",
+        "hover": "optional hover text",
+    },
+    # ...
+]
+
+# t-SNE or PCA projection
+plot = visualize_embeddings(docs=docs, projection="t-sne")  # or "pca"
+
+# Custom per-axis projection
+plot = visualize_embeddings(
+    docs=docs,
+    x_projection=PrincipalComponent(1),
+    y_projection=TextEmbedding("some text to project onto"),
+    title="My Embedding Plot",
+)
+
+plot.display()              # show in Jupyter or browser
+plot.to_html("plot.html")   # self-contained HTML file
+```
+
+## Doc fields
+
+| Field | Required | Description |
+|-------|----------|-------------|
+| `text` | yes | Text to embed |
+| `label` | no | Legend group name; points with the same label share a color |
+| `color` | no | Point color. If `label` is set, applies to the whole group |
+| `line-id` | no | Connects points with the same id in document order |
+| `hover` | no | Custom hover text (defaults to first 100 chars of `text`) |
+
+## Projections
+
+- **`projection="t-sne"`** — t-SNE with cosine metric (default)
+- **`projection="pca"`** — PCA
+- **`PrincipalComponent(n)`** — project onto the nth principal component (1-indexed)
+- **`TextEmbedding("text")`** — cosine similarity with a reference text's embedding
+
+## Example
+
+`examples/repo_files.py` embeds every Python file in this repository at multiple truncation points and connects versions of the same file with lines:
+
+```bash
+uv run python examples/repo_files.py
+```
+
+## Caching
+
+Embeddings are cached to `~/.cache/embedding_visualizer/` so repeated runs don't re-call the API.